Molecular Characterization of Oat Seed Globulins 1 Received for

We have isolated full-length cDNA clones that encode oat (Avena sativa) seed storage globulin mRNAs from a cDNA library in the expression vector lambda gtll. The longest of these clones, pOG2, has an 1840-base pair insert that encodes a complete precursor subunit with a signal peptide of 24 amino acids followed by an acidic polypeptide of 293 amino acids and a basic polypeptide of 201 amino acids. Near the C terminus of the acidic polypeptide are four repeats of a higly conserved, glutamine-rich octapeptide. Other oat globulin cDNA clones contain five of these repeats. Nucleotide sequence comparisons between these clones indicate that the genes encoding these proteins are highly conserved. We estimate there to be 7 to 10 genes for the oat globulin per haploid genome. Comparisons of amino acid sequences show that the oat globulin is 30 to 40% homologous with storage globulins of legumes and about 70% homologous with the rice seed storage globulin (glutelin). During their development, plant seeds accumulate large amounts of storage proteins that serve as sources of nitrogen, sulfur, and carbon compounds during seed germination (25). Two major classes of storage proteins can be distinguished: globulins, which are found principally in the cotyledons and axis of the embryo, and prolamines, which are found in the endosperm of cereal seeds (13). Two major types of storage globulins have been described that have sedimentation coefficients ofabout 7S and 11S. The proportion of the two types of globulins is variable among dicots; monocot embryos contain only the 7S globulin, which is present in the scutellum. Storage globulins account for most of the protein in dicot seeds, but they generally make up only a small fraction of the protein found in cereal seeds. Instead, most cereals contain predominantly prolamine-type storage proteins. Oats and rice are exceptions. These two contain only small amounts ofprolamine (5-10%), and most of their storage protein is an 11 to 12S globulin that is synthesized in the endosperm. Both of these proteins are structurally related to the 11S globulins found in dicots (25), but both are much less soluble than the dicot 11S globulins. The oat globulin requires 0.8 to 1.0 M NaCl for solubility, whereas the characteristics of the rice globulin (glutelin) are such that denaturing solvents are required for solvation. Previous studies in our laboratory (29) and elsewhere (5, 7) 'Supported by a grant from the United States Department of Agriculture Competitive Research Grants Office (87-CRCR-1-2356) to B. A. L. This is Journal Paper No. 1 1,423 of the Purdue Agricultural Research Station. 2Recipient of a postdoctoral fellowship from the National Science Foundation. have partially characterized the structure ofthe oat 12S globulin. This protein is a hexamer of apparent Mr 320,000 (21). The subunits are Mr 54,000, and each consists ofan Mr 32,000 acidic polypeptide that is disulfide bonded to an Mr 22,000 basic polypeptide. We also reported the isolation of a cDNA clone corresponding to one of these subunits that contains the coding sequence for all ofthe basic polypeptide but only a small portion of that for the acidic polypeptide (30). We now report the isolation ofcDNA clones containing the entire coding sequence of the oat 12S globulin. DNA sequence analysis ofthe longest of these cDNA clones shows that it codes for a polypeptide of the size and amino acid composition of purified oat globulin. Structural comparison ofthis protein with those from legumes suggests that the reduced solubility of the oat globulin is due to the presence of repeated peptides of neutral isoelectric point that are found near the C terminus of the acidic polypeptide. We also show that the amino acid sequence predicted by the oat globulin cDNA is more similar to that of the globulin (glutelin) of rice seeds than to the 1 S globulins of legumes. MATERIALS AND METHODS Materials. Restriction endonucleases, EcoRI linkers, and DNA ligase were purchased from Bethesda Research Laboratories (Gaithersburg, MD). [a-32P]dCTP, [a-32P]dATP, and [a-35 ] dATP were purchased from New England Nuclear (Boston, MA) or from Amersham (Arlington Heights, IL). Nick translation kits were purchased from Amersham. Nitrocellulose was from Schleicher and Schuell (Keene, NH). Lambda gtl 1 arms and in vitro packaging extract were obtained from Promega (Madison, WI). Construction of an Oat cDNA Library in X-tll. Total RNA was isolated from frozen immature oat (Avena sativa) grains (10-15 DAF) of cultivar Gary by the procedure of Hall et al. (10). The RNA was suspended in 10 mM Tris-HCl (pH 7.5) and 500 mm KCI, and poly(A)-containing RNA was obtained by one cycle of oligo(dT)-cellulose chromatography (3). After precipitation in ethanol, the poly(A+) RNA was dissolved in sterile water. cDNA was synthesized, treated with EcoRI methylase, and tailed with synthetic EcoRI linkers as described by Huynh et al. (12). The cDNA was size-fractionated by chromatography on a Bio-Gel A-50m column (12), and fractions containingthe largest cDNAs were collected and used in the construction ofthe library. These fractions contained about 0.5 ug ofcDNA of average size 2.2 to 2.5 kb3. One hundred ng of the EcoRI-linked cDNA was ligated with 2 ,g of X-gtl 1 arms and assembled into phage particles in vitro. A total of 2.9 x 101 recombinant clones were obtained, and the library was stored at 4C. Screening the cDNA Library with Oat Globulin Antiserum. A total of 1.2 x 104 phage clones were grown on Eschericia coli. Antiserum against oat globulin (29) was used for serological 3 Abbreviations: kb, kilobase pairs; DPA, days postanthesis.